Oligonucleotides are presently synthesized in vitro using organic synthesis methods. These methods include the phophoramidite method described in Adams et al., J. Amer. Chem. Soc., 105:661 (1983) and Froehler et al., Tetrahedron Lett., 24:3171 (1983) and the phosphotriester method described in German Offenlegungsshrift 264432. Other organic synthesis methods include those described by Froehler et al., U.S. Pat. No. 5,264,566 in which H-phosphonates are used to produce oligonucleotides.
The phosphoramidite method of phosphodiester bond formation and oligonucleotide synthesis represents the current state of the art employed by most laboratories for the coupling of desired nucleotides without the use of a template. In this method, the coupling reaction is initiated by a nucleoside attached to a solid support. The 5'-hydroxyl group of the immobilized nucleoside is free for coupling with the second nucleoside of the chain to be assembled. Since the growing oligonucleotide chain projects a 5'-hydroxyl available for reaction with a mononucleotide, the direction of synthesis if referred to as 3' to 5'.
Each successive mononucleotide to be added to the growing oligonucleotide chain contains a 3' -phosphoramidate moiety which reacts with the 5'-hydroxyl group of the support-bound nucleotide to form a 5' to 3' internucleotide phosphodiester bond. The 5'-hydroxyl group of the incoming mononucleotide is protected, usually by a trityl group, in order to prevent the uncontrolled polymerization of the nucleosides. After each incoming nucleoside is added, the protected 5'-hydroxyl group is deprotected, so that it is available for reaction with the next incoming nucleoside having a 3'-phosphoramidite group and a protected 5'-hydroxyl. This is followed by deprotection and addition of the next incoming nucleotide, and so forth.
Between each nucleoside addition step, unreacted chains which fail to participate in phosphodiester bond formation with the desired nucleoside are chemically "capped" to prevent their further elongation. This usually involves chemical acetylation.
This method and the other currently used organic methods while widely accepted require large amounts of costly monomers that require complex organic synthesis schemes to produce. In addition, these methods are complex in that the phosphoramidite method requires an oxidation step after each condensation reaction. The phosphotriester method requires that the subpopulation of oligonucleotides that have not had a monomer added in a particular cycle be capped in a separate reaction to prevent further chain elongation of these oligonucleotides.
Other drawbacks of virtually all chemical methods of phosphodiester bond formation, is that the reaction must be performed in organic solvents and in the absence of water. Many of these organic solvents are toxic or otherwise hazardous. Another drawback to chemical synthesis is that it is at best 98 percent efficient at each cycle. In other words, following each nucleotide addition, at least 2 percent of the growing oligonucleotide chains are capped, resulting in a yield loss. The total yield loss for the nucleotide chain being synthesized thus increases with each nucleotide added to the sequence.
For example, assuming a yield of 98 percent per nucleotide addition, the synthesis of a polynucleotide consisting of 70 mononucleotides would experience a yield loss of nearly 75 percent. Furthermore, the object nucleotide chain would require isolation from a reaction mixture of polynucleotides, nearly 75 percent of which consist of capped oligonucleotides ranging between 1 and 69 nucleotides in length.
This inherent inefficiency in chemical synthesis of oligonucleotides ultimately limits the length of oligonucleotide that can be efficiently produced to oligonucleotides having 50 nucleic acid residues or less.
A need exists for a method which improves the efficiency of phosphodiester bond formation and which could ultimately be capable of producing shorter chain oligonucleotides in higher yields and longer chain polynucleotides in acceptable yields. In addition, a need exists for a polynucleotide synthesis system which is compatible with pre-existing polynucleotides, such as vector DNAs, so that desired polynucleotide sequences can readily be added on to the pre-existing sequences. Chemical coupling by the phosphoramidite method is not compatible with "add-on" synthesis to pre-existing polynucleotides. Enzyme catalyzed phosphodiester bond formation, however, can be performed in an aqueous environment utilizing either single or double stranded oligo- or polynucleotides to initiate the reaction. These reaction conditions also greatly minimize the use of toxic and hazardous materials.
The 3' to 5' direction of synthesis inherent to the phosphoramidite method of phosphodiester bond formation cannot be enzyme catalyzed. All known enzymes capable of catalyzing the formation of phosphodiester bonds do so in the 5' to 3' direction since the growing polynucleotide strand always projects a 3'-hydroxyl available for attachment of the next nucleoside.
There are many enzymes capable of catalyzing the formation of phosphodiester bonds. One class of such enzymes, the polymerases, are largely template dependent in that they add a complementary nucleotide to the 3' hydroxyl of the growing strand of a double stranded polynucleotide. However, some polymerases are template independent and primarily catalyze the formation of single stranded nucleotide polymers. Another class of enzyme, the ligases, are template independent and form a phosphodiester bond between two polynucleotides or between a polynucleotide and a mononucleotide.
Addition of single nucleotides to DNA fragments, catalyzed by deoxynucleotidyl terminal transferase (TdTase), has previously been described by Deng and Wu, Meth. Enzymol., 100:96-116, 1983. These reaction conditions did not involve transient protection of the 3'-hydroxyl nor were they intended to be used for the sequential creation of phosphodiester bonds to synthesize a predetermined nucleotide sequence. The presence of unprotected 3'-hydroxyls resulted in a highly heterogeneous population of reaction products.
Similarly, prior attempts to catalyze synthesis of very short pieces of RNA or DNA using protected nucleotide monophosphates or diphosphates resulted in unacceptably low levels of the desired phosphodiester bond formation or required excessive amounts of enzyme to achieve acceptable efficiencies. These problems were largely due to unavoidable heterogeneity of the mononucleotide building blocks or to the very high turnover number of the enzyme, necessitating extremely long incubation times (see, for example, Hinton and Gumport, Nucleic Acids Res. 7:453-464, 1979; Kaufman et al., Eur. J. Biochem., 24: 4-11, 1971). These experiments were limited to 5'-monophosphates and diphosphates. No attempts have been made to catalyze controlled DNA synthesis using 5'-triphosphates protected at the 3' position.
Enzyme catalyzed creation of a single phosphodiester bond between the free 3'-hydroxyl group of an oligonucleotide chain and the 5'-phosphate of a mononucleotide requires protection of the 3'-hydroxyl of the mononucleotide in order to prevent multiple phosphodiester bond formations and hence repeated mononucleotide additions. Protection of the 3'-hydroxyl of the mononucleotide ideally involves a transient blocking group which can readily be removed in order to allow subsequent reactions. Flugel et al., Biochem. Biophys. Acta. 308:35-40, 1973, report that nucleoside triphosphates with blocked 3'-hydroxyl groups cannot be prepared directly. This lack of 3' blocked triphosphates necessitated previous processes to utilize lower energy and thus more inefficient 3' blocked monophosphates and diposphates. Synthetic techniques to create 3' block triphosphates would be highly desirable, because this could enable stepwise enzyme catalyzed phosphodiester bond formation leading to polynucleotide synthesis.
These prior attempts at synthesizing oligonucleotides using a template independent polymerase were extremely inefficient resulting in the production of very short oligonucleotides. The inefficiency of these methods made these methods useless for practical synthesis of oligonucleotides.
The present invention allows the creation of phosphodiester bonds between nucleotides using a template independent polymerase to create oligonucleotides having a predetermined sequence. This enzyme catalysis can vastly improve the efficiency of phosphodiester bond formation between desired nucleotides compared to current techniques of chemical coupling and can be carried out in the presence of other biological molecules such as pre-existing sequences of single or double stranded DNA as well as other types of enzymes. In addition, the very high specificity inherent to enzyme catalysis allows only coupling of a 5'-phosphate to a 3'-hydroxyl. The coupling of two mononucleosides, as well as various other side reactions inherent to chemical coupling techniques, simply do not occur.
A further advantage of the present invention is realized by using 3' blocked triphosphates having high energy phosphate bonds which an enzyme can utilize to drive the reaction to greater completion level than when other monophosphates and diphosphates are used. In addition, triphosphates are less strongly hydrated than the diphosphate, which also tends to drive catalytic hydrolysis of the triphosphate to completion.
Clearly, the availability of a homogeneous population of protected mononucleotide triphosphates and enzymes capable of efficiently joining protected nucleotides to initiating substrates will enable the creation of a highly uniform population of synthetic polynucleotides resulting from stepwise phosphodiester bond formation.